kern_proc.c revision 1.187
1 /* $NetBSD: kern_proc.c,v 1.187 2013/06/10 14:53:52 pooka Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2006, 2007, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, and by Andrew Doran. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1982, 1986, 1989, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. Neither the name of the University nor the names of its contributors 46 * may be used to endorse or promote products derived from this software 47 * without specific prior written permission. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59 * SUCH DAMAGE. 60 * 61 * @(#)kern_proc.c 8.7 (Berkeley) 2/14/95 62 */ 63 64 #include <sys/cdefs.h> 65 __KERNEL_RCSID(0, "$NetBSD: kern_proc.c,v 1.187 2013/06/10 14:53:52 pooka Exp $"); 66 67 #ifdef _KERNEL_OPT 68 #include "opt_kstack.h" 69 #include "opt_maxuprc.h" 70 #include "opt_dtrace.h" 71 #include "opt_compat_netbsd32.h" 72 #endif 73 74 #include <sys/param.h> 75 #include <sys/systm.h> 76 #include <sys/kernel.h> 77 #include <sys/proc.h> 78 #include <sys/resourcevar.h> 79 #include <sys/buf.h> 80 #include <sys/acct.h> 81 #include <sys/wait.h> 82 #include <sys/file.h> 83 #include <ufs/ufs/quota.h> 84 #include <sys/uio.h> 85 #include <sys/pool.h> 86 #include <sys/pset.h> 87 #include <sys/mbuf.h> 88 #include <sys/ioctl.h> 89 #include <sys/tty.h> 90 #include <sys/signalvar.h> 91 #include <sys/ras.h> 92 #include <sys/filedesc.h> 93 #include <sys/syscall_stats.h> 94 #include <sys/kauth.h> 95 #include <sys/sleepq.h> 96 #include <sys/atomic.h> 97 #include <sys/kmem.h> 98 #include <sys/dtrace_bsd.h> 99 #include <sys/sysctl.h> 100 #include <sys/exec.h> 101 #include <sys/cpu.h> 102 103 #include <uvm/uvm_extern.h> 104 105 #ifdef COMPAT_NETBSD32 106 #include <compat/netbsd32/netbsd32.h> 107 #endif 108 109 /* 110 * Process lists. 111 */ 112 113 struct proclist allproc __cacheline_aligned; 114 struct proclist zombproc __cacheline_aligned; 115 116 kmutex_t * proc_lock __cacheline_aligned; 117 118 /* 119 * pid to proc lookup is done by indexing the pid_table array. 120 * Since pid numbers are only allocated when an empty slot 121 * has been found, there is no need to search any lists ever. 122 * (an orphaned pgrp will lock the slot, a session will lock 123 * the pgrp with the same number.) 124 * If the table is too small it is reallocated with twice the 125 * previous size and the entries 'unzipped' into the two halves. 126 * A linked list of free entries is passed through the pt_proc 127 * field of 'free' items - set odd to be an invalid ptr. 128 */ 129 130 struct pid_table { 131 struct proc *pt_proc; 132 struct pgrp *pt_pgrp; 133 pid_t pt_pid; 134 }; 135 #if 1 /* strongly typed cast - should be a noop */ 136 static inline uint p2u(struct proc *p) { return (uint)(uintptr_t)p; } 137 #else 138 #define p2u(p) ((uint)p) 139 #endif 140 #define P_VALID(p) (!(p2u(p) & 1)) 141 #define P_NEXT(p) (p2u(p) >> 1) 142 #define P_FREE(pid) ((struct proc *)(uintptr_t)((pid) << 1 | 1)) 143 144 /* 145 * Table of process IDs (PIDs). 146 */ 147 static struct pid_table *pid_table __read_mostly; 148 149 #define INITIAL_PID_TABLE_SIZE (1 << 5) 150 151 /* Table mask, threshold for growing and number of allocated PIDs. */ 152 static u_int pid_tbl_mask __read_mostly; 153 static u_int pid_alloc_lim __read_mostly; 154 static u_int pid_alloc_cnt __cacheline_aligned; 155 156 /* Next free, last free and maximum PIDs. */ 157 static u_int next_free_pt __cacheline_aligned; 158 static u_int last_free_pt __cacheline_aligned; 159 static pid_t pid_max __read_mostly; 160 161 /* Components of the first process -- never freed. */ 162 163 extern struct emul emul_netbsd; /* defined in kern_exec.c */ 164 165 struct session session0 = { 166 .s_count = 1, 167 .s_sid = 0, 168 }; 169 struct pgrp pgrp0 = { 170 .pg_members = LIST_HEAD_INITIALIZER(&pgrp0.pg_members), 171 .pg_session = &session0, 172 }; 173 filedesc_t filedesc0; 174 struct cwdinfo cwdi0 = { 175 .cwdi_cmask = CMASK, 176 .cwdi_refcnt = 1, 177 }; 178 struct plimit limit0; 179 struct pstats pstat0; 180 struct vmspace vmspace0; 181 struct sigacts sigacts0; 182 struct proc proc0 = { 183 .p_lwps = LIST_HEAD_INITIALIZER(&proc0.p_lwps), 184 .p_sigwaiters = LIST_HEAD_INITIALIZER(&proc0.p_sigwaiters), 185 .p_nlwps = 1, 186 .p_nrlwps = 1, 187 .p_nlwpid = 1, /* must match lwp0.l_lid */ 188 .p_pgrp = &pgrp0, 189 .p_comm = "system", 190 /* 191 * Set P_NOCLDWAIT so that kernel threads are reparented to init(8) 192 * when they exit. init(8) can easily wait them out for us. 193 */ 194 .p_flag = PK_SYSTEM | PK_NOCLDWAIT, 195 .p_stat = SACTIVE, 196 .p_nice = NZERO, 197 .p_emul = &emul_netbsd, 198 .p_cwdi = &cwdi0, 199 .p_limit = &limit0, 200 .p_fd = &filedesc0, 201 .p_vmspace = &vmspace0, 202 .p_stats = &pstat0, 203 .p_sigacts = &sigacts0, 204 }; 205 kauth_cred_t cred0; 206 207 static const int nofile = NOFILE; 208 static const int maxuprc = MAXUPRC; 209 210 static int sysctl_doeproc(SYSCTLFN_PROTO); 211 static int sysctl_kern_proc_args(SYSCTLFN_PROTO); 212 static void fill_kproc2(struct proc *, struct kinfo_proc2 *, bool); 213 214 /* 215 * The process list descriptors, used during pid allocation and 216 * by sysctl. No locking on this data structure is needed since 217 * it is completely static. 218 */ 219 const struct proclist_desc proclists[] = { 220 { &allproc }, 221 { &zombproc }, 222 { NULL }, 223 }; 224 225 static struct pgrp * pg_remove(pid_t); 226 static void pg_delete(pid_t); 227 static void orphanpg(struct pgrp *); 228 229 static specificdata_domain_t proc_specificdata_domain; 230 231 static pool_cache_t proc_cache; 232 233 static kauth_listener_t proc_listener; 234 235 static int 236 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 237 void *arg0, void *arg1, void *arg2, void *arg3) 238 { 239 struct proc *p; 240 int result; 241 242 result = KAUTH_RESULT_DEFER; 243 p = arg0; 244 245 switch (action) { 246 case KAUTH_PROCESS_CANSEE: { 247 enum kauth_process_req req; 248 249 req = (enum kauth_process_req)arg1; 250 251 switch (req) { 252 case KAUTH_REQ_PROCESS_CANSEE_ARGS: 253 case KAUTH_REQ_PROCESS_CANSEE_ENTRY: 254 case KAUTH_REQ_PROCESS_CANSEE_OPENFILES: 255 result = KAUTH_RESULT_ALLOW; 256 257 break; 258 259 case KAUTH_REQ_PROCESS_CANSEE_ENV: 260 if (kauth_cred_getuid(cred) != 261 kauth_cred_getuid(p->p_cred) || 262 kauth_cred_getuid(cred) != 263 kauth_cred_getsvuid(p->p_cred)) 264 break; 265 266 result = KAUTH_RESULT_ALLOW; 267 268 break; 269 270 default: 271 break; 272 } 273 274 break; 275 } 276 277 case KAUTH_PROCESS_FORK: { 278 int lnprocs = (int)(unsigned long)arg2; 279 280 /* 281 * Don't allow a nonprivileged user to use the last few 282 * processes. The variable lnprocs is the current number of 283 * processes, maxproc is the limit. 284 */ 285 if (__predict_false((lnprocs >= maxproc - 5))) 286 break; 287 288 result = KAUTH_RESULT_ALLOW; 289 290 break; 291 } 292 293 case KAUTH_PROCESS_CORENAME: 294 case KAUTH_PROCESS_STOPFLAG: 295 if (proc_uidmatch(cred, p->p_cred) == 0) 296 result = KAUTH_RESULT_ALLOW; 297 298 break; 299 300 default: 301 break; 302 } 303 304 return result; 305 } 306 307 /* 308 * Initialize global process hashing structures. 309 */ 310 void 311 procinit(void) 312 { 313 const struct proclist_desc *pd; 314 u_int i; 315 #define LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1)) 316 317 for (pd = proclists; pd->pd_list != NULL; pd++) 318 LIST_INIT(pd->pd_list); 319 320 proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 321 pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE 322 * sizeof(struct pid_table), KM_SLEEP); 323 pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1; 324 pid_max = PID_MAX; 325 326 /* Set free list running through table... 327 Preset 'use count' above PID_MAX so we allocate pid 1 next. */ 328 for (i = 0; i <= pid_tbl_mask; i++) { 329 pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1); 330 pid_table[i].pt_pgrp = 0; 331 pid_table[i].pt_pid = 0; 332 } 333 /* slot 0 is just grabbed */ 334 next_free_pt = 1; 335 /* Need to fix last entry. */ 336 last_free_pt = pid_tbl_mask; 337 pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY); 338 /* point at which we grow table - to avoid reusing pids too often */ 339 pid_alloc_lim = pid_tbl_mask - 1; 340 #undef LINK_EMPTY 341 342 proc_specificdata_domain = specificdata_domain_create(); 343 KASSERT(proc_specificdata_domain != NULL); 344 345 proc_cache = pool_cache_init(sizeof(struct proc), 0, 0, 0, 346 "procpl", NULL, IPL_NONE, NULL, NULL, NULL); 347 348 proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS, 349 proc_listener_cb, NULL); 350 } 351 352 void 353 procinit_sysctl(void) 354 { 355 static struct sysctllog *clog; 356 357 sysctl_createv(&clog, 0, NULL, NULL, 358 CTLFLAG_PERMANENT, 359 CTLTYPE_NODE, "kern", NULL, 360 NULL, 0, NULL, 0, 361 CTL_KERN, CTL_EOL); 362 363 sysctl_createv(&clog, 0, NULL, NULL, 364 CTLFLAG_PERMANENT, 365 CTLTYPE_NODE, "proc", 366 SYSCTL_DESCR("System-wide process information"), 367 sysctl_doeproc, 0, NULL, 0, 368 CTL_KERN, KERN_PROC, CTL_EOL); 369 sysctl_createv(&clog, 0, NULL, NULL, 370 CTLFLAG_PERMANENT, 371 CTLTYPE_NODE, "proc2", 372 SYSCTL_DESCR("Machine-independent process information"), 373 sysctl_doeproc, 0, NULL, 0, 374 CTL_KERN, KERN_PROC2, CTL_EOL); 375 sysctl_createv(&clog, 0, NULL, NULL, 376 CTLFLAG_PERMANENT, 377 CTLTYPE_NODE, "proc_args", 378 SYSCTL_DESCR("Process argument information"), 379 sysctl_kern_proc_args, 0, NULL, 0, 380 CTL_KERN, KERN_PROC_ARGS, CTL_EOL); 381 382 /* 383 "nodes" under these: 384 385 KERN_PROC_ALL 386 KERN_PROC_PID pid 387 KERN_PROC_PGRP pgrp 388 KERN_PROC_SESSION sess 389 KERN_PROC_TTY tty 390 KERN_PROC_UID uid 391 KERN_PROC_RUID uid 392 KERN_PROC_GID gid 393 KERN_PROC_RGID gid 394 395 all in all, probably not worth the effort... 396 */ 397 } 398 399 /* 400 * Initialize process 0. 401 */ 402 void 403 proc0_init(void) 404 { 405 struct proc *p; 406 struct pgrp *pg; 407 struct rlimit *rlim; 408 rlim_t lim; 409 int i; 410 411 p = &proc0; 412 pg = &pgrp0; 413 414 mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 415 mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 416 p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 417 418 rw_init(&p->p_reflock); 419 cv_init(&p->p_waitcv, "wait"); 420 cv_init(&p->p_lwpcv, "lwpwait"); 421 422 LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling); 423 424 pid_table[0].pt_proc = p; 425 LIST_INSERT_HEAD(&allproc, p, p_list); 426 427 pid_table[0].pt_pgrp = pg; 428 LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist); 429 430 #ifdef __HAVE_SYSCALL_INTERN 431 (*p->p_emul->e_syscall_intern)(p); 432 #endif 433 434 /* Create credentials. */ 435 cred0 = kauth_cred_alloc(); 436 p->p_cred = cred0; 437 438 /* Create the CWD info. */ 439 rw_init(&cwdi0.cwdi_lock); 440 441 /* Create the limits structures. */ 442 mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE); 443 444 rlim = limit0.pl_rlimit; 445 for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) { 446 rlim[i].rlim_cur = RLIM_INFINITY; 447 rlim[i].rlim_max = RLIM_INFINITY; 448 } 449 450 rlim[RLIMIT_NOFILE].rlim_max = maxfiles; 451 rlim[RLIMIT_NOFILE].rlim_cur = maxfiles < nofile ? maxfiles : nofile; 452 453 rlim[RLIMIT_NPROC].rlim_max = maxproc; 454 rlim[RLIMIT_NPROC].rlim_cur = maxproc < maxuprc ? maxproc : maxuprc; 455 456 lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvmexp.free)); 457 rlim[RLIMIT_RSS].rlim_max = lim; 458 rlim[RLIMIT_MEMLOCK].rlim_max = lim; 459 rlim[RLIMIT_MEMLOCK].rlim_cur = lim / 3; 460 461 rlim[RLIMIT_NTHR].rlim_max = maxlwp; 462 rlim[RLIMIT_NTHR].rlim_cur = maxlwp < maxuprc ? maxlwp : maxuprc; 463 464 /* Note that default core name has zero length. */ 465 limit0.pl_corename = defcorename; 466 limit0.pl_cnlen = 0; 467 limit0.pl_refcnt = 1; 468 limit0.pl_writeable = false; 469 limit0.pl_sv_limit = NULL; 470 471 /* Configure virtual memory system, set vm rlimits. */ 472 uvm_init_limits(p); 473 474 /* Initialize file descriptor table for proc0. */ 475 fd_init(&filedesc0); 476 477 /* 478 * Initialize proc0's vmspace, which uses the kernel pmap. 479 * All kernel processes (which never have user space mappings) 480 * share proc0's vmspace, and thus, the kernel pmap. 481 */ 482 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 483 trunc_page(VM_MAX_ADDRESS)); 484 485 /* Initialize signal state for proc0. XXX IPL_SCHED */ 486 mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED); 487 siginit(p); 488 489 proc_initspecific(p); 490 kdtrace_proc_ctor(NULL, p); 491 } 492 493 /* 494 * Session reference counting. 495 */ 496 497 void 498 proc_sesshold(struct session *ss) 499 { 500 501 KASSERT(mutex_owned(proc_lock)); 502 ss->s_count++; 503 } 504 505 void 506 proc_sessrele(struct session *ss) 507 { 508 509 KASSERT(mutex_owned(proc_lock)); 510 /* 511 * We keep the pgrp with the same id as the session in order to 512 * stop a process being given the same pid. Since the pgrp holds 513 * a reference to the session, it must be a 'zombie' pgrp by now. 514 */ 515 if (--ss->s_count == 0) { 516 struct pgrp *pg; 517 518 pg = pg_remove(ss->s_sid); 519 mutex_exit(proc_lock); 520 521 kmem_free(pg, sizeof(struct pgrp)); 522 kmem_free(ss, sizeof(struct session)); 523 } else { 524 mutex_exit(proc_lock); 525 } 526 } 527 528 /* 529 * Check that the specified process group is in the session of the 530 * specified process. 531 * Treats -ve ids as process ids. 532 * Used to validate TIOCSPGRP requests. 533 */ 534 int 535 pgid_in_session(struct proc *p, pid_t pg_id) 536 { 537 struct pgrp *pgrp; 538 struct session *session; 539 int error; 540 541 mutex_enter(proc_lock); 542 if (pg_id < 0) { 543 struct proc *p1 = proc_find(-pg_id); 544 if (p1 == NULL) { 545 error = EINVAL; 546 goto fail; 547 } 548 pgrp = p1->p_pgrp; 549 } else { 550 pgrp = pgrp_find(pg_id); 551 if (pgrp == NULL) { 552 error = EINVAL; 553 goto fail; 554 } 555 } 556 session = pgrp->pg_session; 557 error = (session != p->p_pgrp->pg_session) ? EPERM : 0; 558 fail: 559 mutex_exit(proc_lock); 560 return error; 561 } 562 563 /* 564 * p_inferior: is p an inferior of q? 565 */ 566 static inline bool 567 p_inferior(struct proc *p, struct proc *q) 568 { 569 570 KASSERT(mutex_owned(proc_lock)); 571 572 for (; p != q; p = p->p_pptr) 573 if (p->p_pid == 0) 574 return false; 575 return true; 576 } 577 578 /* 579 * proc_find: locate a process by the ID. 580 * 581 * => Must be called with proc_lock held. 582 */ 583 proc_t * 584 proc_find_raw(pid_t pid) 585 { 586 struct pid_table *pt; 587 proc_t *p; 588 589 KASSERT(mutex_owned(proc_lock)); 590 pt = &pid_table[pid & pid_tbl_mask]; 591 p = pt->pt_proc; 592 if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) { 593 return NULL; 594 } 595 return p; 596 } 597 598 proc_t * 599 proc_find(pid_t pid) 600 { 601 proc_t *p; 602 603 p = proc_find_raw(pid); 604 if (__predict_false(p == NULL)) { 605 return NULL; 606 } 607 608 /* 609 * Only allow live processes to be found by PID. 610 * XXX: p_stat might change, since unlocked. 611 */ 612 if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) { 613 return p; 614 } 615 return NULL; 616 } 617 618 /* 619 * pgrp_find: locate a process group by the ID. 620 * 621 * => Must be called with proc_lock held. 622 */ 623 struct pgrp * 624 pgrp_find(pid_t pgid) 625 { 626 struct pgrp *pg; 627 628 KASSERT(mutex_owned(proc_lock)); 629 630 pg = pid_table[pgid & pid_tbl_mask].pt_pgrp; 631 632 /* 633 * Cannot look up a process group that only exists because the 634 * session has not died yet (traditional). 635 */ 636 if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) { 637 return NULL; 638 } 639 return pg; 640 } 641 642 static void 643 expand_pid_table(void) 644 { 645 size_t pt_size, tsz; 646 struct pid_table *n_pt, *new_pt; 647 struct proc *proc; 648 struct pgrp *pgrp; 649 pid_t pid, rpid; 650 u_int i; 651 uint new_pt_mask; 652 653 pt_size = pid_tbl_mask + 1; 654 tsz = pt_size * 2 * sizeof(struct pid_table); 655 new_pt = kmem_alloc(tsz, KM_SLEEP); 656 new_pt_mask = pt_size * 2 - 1; 657 658 mutex_enter(proc_lock); 659 if (pt_size != pid_tbl_mask + 1) { 660 /* Another process beat us to it... */ 661 mutex_exit(proc_lock); 662 kmem_free(new_pt, tsz); 663 return; 664 } 665 666 /* 667 * Copy entries from old table into new one. 668 * If 'pid' is 'odd' we need to place in the upper half, 669 * even pid's to the lower half. 670 * Free items stay in the low half so we don't have to 671 * fixup the reference to them. 672 * We stuff free items on the front of the freelist 673 * because we can't write to unmodified entries. 674 * Processing the table backwards maintains a semblance 675 * of issuing pid numbers that increase with time. 676 */ 677 i = pt_size - 1; 678 n_pt = new_pt + i; 679 for (; ; i--, n_pt--) { 680 proc = pid_table[i].pt_proc; 681 pgrp = pid_table[i].pt_pgrp; 682 if (!P_VALID(proc)) { 683 /* Up 'use count' so that link is valid */ 684 pid = (P_NEXT(proc) + pt_size) & ~pt_size; 685 rpid = 0; 686 proc = P_FREE(pid); 687 if (pgrp) 688 pid = pgrp->pg_id; 689 } else { 690 pid = pid_table[i].pt_pid; 691 rpid = pid; 692 } 693 694 /* Save entry in appropriate half of table */ 695 n_pt[pid & pt_size].pt_proc = proc; 696 n_pt[pid & pt_size].pt_pgrp = pgrp; 697 n_pt[pid & pt_size].pt_pid = rpid; 698 699 /* Put other piece on start of free list */ 700 pid = (pid ^ pt_size) & ~pid_tbl_mask; 701 n_pt[pid & pt_size].pt_proc = 702 P_FREE((pid & ~pt_size) | next_free_pt); 703 n_pt[pid & pt_size].pt_pgrp = 0; 704 n_pt[pid & pt_size].pt_pid = 0; 705 706 next_free_pt = i | (pid & pt_size); 707 if (i == 0) 708 break; 709 } 710 711 /* Save old table size and switch tables */ 712 tsz = pt_size * sizeof(struct pid_table); 713 n_pt = pid_table; 714 pid_table = new_pt; 715 pid_tbl_mask = new_pt_mask; 716 717 /* 718 * pid_max starts as PID_MAX (= 30000), once we have 16384 719 * allocated pids we need it to be larger! 720 */ 721 if (pid_tbl_mask > PID_MAX) { 722 pid_max = pid_tbl_mask * 2 + 1; 723 pid_alloc_lim |= pid_alloc_lim << 1; 724 } else 725 pid_alloc_lim <<= 1; /* doubles number of free slots... */ 726 727 mutex_exit(proc_lock); 728 kmem_free(n_pt, tsz); 729 } 730 731 struct proc * 732 proc_alloc(void) 733 { 734 struct proc *p; 735 736 p = pool_cache_get(proc_cache, PR_WAITOK); 737 p->p_stat = SIDL; /* protect against others */ 738 proc_initspecific(p); 739 kdtrace_proc_ctor(NULL, p); 740 p->p_pid = -1; 741 proc_alloc_pid(p); 742 return p; 743 } 744 745 /* 746 * proc_alloc_pid: allocate PID and record the given proc 'p' so that 747 * proc_find_raw() can find it by the PID. 748 */ 749 750 pid_t 751 proc_alloc_pid(struct proc *p) 752 { 753 struct pid_table *pt; 754 pid_t pid; 755 int nxt; 756 757 for (;;expand_pid_table()) { 758 if (__predict_false(pid_alloc_cnt >= pid_alloc_lim)) 759 /* ensure pids cycle through 2000+ values */ 760 continue; 761 mutex_enter(proc_lock); 762 pt = &pid_table[next_free_pt]; 763 #ifdef DIAGNOSTIC 764 if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp)) 765 panic("proc_alloc: slot busy"); 766 #endif 767 nxt = P_NEXT(pt->pt_proc); 768 if (nxt & pid_tbl_mask) 769 break; 770 /* Table full - expand (NB last entry not used....) */ 771 mutex_exit(proc_lock); 772 } 773 774 /* pid is 'saved use count' + 'size' + entry */ 775 pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt; 776 if ((uint)pid > (uint)pid_max) 777 pid &= pid_tbl_mask; 778 next_free_pt = nxt & pid_tbl_mask; 779 780 /* Grab table slot */ 781 pt->pt_proc = p; 782 783 KASSERT(pt->pt_pid == 0); 784 pt->pt_pid = pid; 785 if (p->p_pid == -1) { 786 p->p_pid = pid; 787 } 788 pid_alloc_cnt++; 789 mutex_exit(proc_lock); 790 791 return pid; 792 } 793 794 /* 795 * Free a process id - called from proc_free (in kern_exit.c) 796 * 797 * Called with the proc_lock held. 798 */ 799 void 800 proc_free_pid(pid_t pid) 801 { 802 struct pid_table *pt; 803 804 KASSERT(mutex_owned(proc_lock)); 805 806 pt = &pid_table[pid & pid_tbl_mask]; 807 808 /* save pid use count in slot */ 809 pt->pt_proc = P_FREE(pid & ~pid_tbl_mask); 810 KASSERT(pt->pt_pid == pid); 811 pt->pt_pid = 0; 812 813 if (pt->pt_pgrp == NULL) { 814 /* link last freed entry onto ours */ 815 pid &= pid_tbl_mask; 816 pt = &pid_table[last_free_pt]; 817 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid); 818 pt->pt_pid = 0; 819 last_free_pt = pid; 820 pid_alloc_cnt--; 821 } 822 823 atomic_dec_uint(&nprocs); 824 } 825 826 void 827 proc_free_mem(struct proc *p) 828 { 829 830 kdtrace_proc_dtor(NULL, p); 831 pool_cache_put(proc_cache, p); 832 } 833 834 /* 835 * proc_enterpgrp: move p to a new or existing process group (and session). 836 * 837 * If we are creating a new pgrp, the pgid should equal 838 * the calling process' pid. 839 * If is only valid to enter a process group that is in the session 840 * of the process. 841 * Also mksess should only be set if we are creating a process group 842 * 843 * Only called from sys_setsid, sys_setpgid and posix_spawn/spawn_return. 844 */ 845 int 846 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess) 847 { 848 struct pgrp *new_pgrp, *pgrp; 849 struct session *sess; 850 struct proc *p; 851 int rval; 852 pid_t pg_id = NO_PGID; 853 854 sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL; 855 856 /* Allocate data areas we might need before doing any validity checks */ 857 mutex_enter(proc_lock); /* Because pid_table might change */ 858 if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) { 859 mutex_exit(proc_lock); 860 new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP); 861 mutex_enter(proc_lock); 862 } else 863 new_pgrp = NULL; 864 rval = EPERM; /* most common error (to save typing) */ 865 866 /* Check pgrp exists or can be created */ 867 pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp; 868 if (pgrp != NULL && pgrp->pg_id != pgid) 869 goto done; 870 871 /* Can only set another process under restricted circumstances. */ 872 if (pid != curp->p_pid) { 873 /* Must exist and be one of our children... */ 874 p = proc_find(pid); 875 if (p == NULL || !p_inferior(p, curp)) { 876 rval = ESRCH; 877 goto done; 878 } 879 /* ... in the same session... */ 880 if (sess != NULL || p->p_session != curp->p_session) 881 goto done; 882 /* ... existing pgid must be in same session ... */ 883 if (pgrp != NULL && pgrp->pg_session != p->p_session) 884 goto done; 885 /* ... and not done an exec. */ 886 if (p->p_flag & PK_EXEC) { 887 rval = EACCES; 888 goto done; 889 } 890 } else { 891 /* ... setsid() cannot re-enter a pgrp */ 892 if (mksess && (curp->p_pgid == curp->p_pid || 893 pgrp_find(curp->p_pid))) 894 goto done; 895 p = curp; 896 } 897 898 /* Changing the process group/session of a session 899 leader is definitely off limits. */ 900 if (SESS_LEADER(p)) { 901 if (sess == NULL && p->p_pgrp == pgrp) 902 /* unless it's a definite noop */ 903 rval = 0; 904 goto done; 905 } 906 907 /* Can only create a process group with id of process */ 908 if (pgrp == NULL && pgid != pid) 909 goto done; 910 911 /* Can only create a session if creating pgrp */ 912 if (sess != NULL && pgrp != NULL) 913 goto done; 914 915 /* Check we allocated memory for a pgrp... */ 916 if (pgrp == NULL && new_pgrp == NULL) 917 goto done; 918 919 /* Don't attach to 'zombie' pgrp */ 920 if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members)) 921 goto done; 922 923 /* Expect to succeed now */ 924 rval = 0; 925 926 if (pgrp == p->p_pgrp) 927 /* nothing to do */ 928 goto done; 929 930 /* Ok all setup, link up required structures */ 931 932 if (pgrp == NULL) { 933 pgrp = new_pgrp; 934 new_pgrp = NULL; 935 if (sess != NULL) { 936 sess->s_sid = p->p_pid; 937 sess->s_leader = p; 938 sess->s_count = 1; 939 sess->s_ttyvp = NULL; 940 sess->s_ttyp = NULL; 941 sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET; 942 memcpy(sess->s_login, p->p_session->s_login, 943 sizeof(sess->s_login)); 944 p->p_lflag &= ~PL_CONTROLT; 945 } else { 946 sess = p->p_pgrp->pg_session; 947 proc_sesshold(sess); 948 } 949 pgrp->pg_session = sess; 950 sess = NULL; 951 952 pgrp->pg_id = pgid; 953 LIST_INIT(&pgrp->pg_members); 954 #ifdef DIAGNOSTIC 955 if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp)) 956 panic("enterpgrp: pgrp table slot in use"); 957 if (__predict_false(mksess && p != curp)) 958 panic("enterpgrp: mksession and p != curproc"); 959 #endif 960 pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp; 961 pgrp->pg_jobc = 0; 962 } 963 964 /* 965 * Adjust eligibility of affected pgrps to participate in job control. 966 * Increment eligibility counts before decrementing, otherwise we 967 * could reach 0 spuriously during the first call. 968 */ 969 fixjobc(p, pgrp, 1); 970 fixjobc(p, p->p_pgrp, 0); 971 972 /* Interlock with ttread(). */ 973 mutex_spin_enter(&tty_lock); 974 975 /* Move process to requested group. */ 976 LIST_REMOVE(p, p_pglist); 977 if (LIST_EMPTY(&p->p_pgrp->pg_members)) 978 /* defer delete until we've dumped the lock */ 979 pg_id = p->p_pgrp->pg_id; 980 p->p_pgrp = pgrp; 981 LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist); 982 983 /* Done with the swap; we can release the tty mutex. */ 984 mutex_spin_exit(&tty_lock); 985 986 done: 987 if (pg_id != NO_PGID) { 988 /* Releases proc_lock. */ 989 pg_delete(pg_id); 990 } else { 991 mutex_exit(proc_lock); 992 } 993 if (sess != NULL) 994 kmem_free(sess, sizeof(*sess)); 995 if (new_pgrp != NULL) 996 kmem_free(new_pgrp, sizeof(*new_pgrp)); 997 #ifdef DEBUG_PGRP 998 if (__predict_false(rval)) 999 printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n", 1000 pid, pgid, mksess, curp->p_pid, rval); 1001 #endif 1002 return rval; 1003 } 1004 1005 /* 1006 * proc_leavepgrp: remove a process from its process group. 1007 * => must be called with the proc_lock held, which will be released; 1008 */ 1009 void 1010 proc_leavepgrp(struct proc *p) 1011 { 1012 struct pgrp *pgrp; 1013 1014 KASSERT(mutex_owned(proc_lock)); 1015 1016 /* Interlock with ttread() */ 1017 mutex_spin_enter(&tty_lock); 1018 pgrp = p->p_pgrp; 1019 LIST_REMOVE(p, p_pglist); 1020 p->p_pgrp = NULL; 1021 mutex_spin_exit(&tty_lock); 1022 1023 if (LIST_EMPTY(&pgrp->pg_members)) { 1024 /* Releases proc_lock. */ 1025 pg_delete(pgrp->pg_id); 1026 } else { 1027 mutex_exit(proc_lock); 1028 } 1029 } 1030 1031 /* 1032 * pg_remove: remove a process group from the table. 1033 * => must be called with the proc_lock held; 1034 * => returns process group to free; 1035 */ 1036 static struct pgrp * 1037 pg_remove(pid_t pg_id) 1038 { 1039 struct pgrp *pgrp; 1040 struct pid_table *pt; 1041 1042 KASSERT(mutex_owned(proc_lock)); 1043 1044 pt = &pid_table[pg_id & pid_tbl_mask]; 1045 pgrp = pt->pt_pgrp; 1046 1047 KASSERT(pgrp != NULL); 1048 KASSERT(pgrp->pg_id == pg_id); 1049 KASSERT(LIST_EMPTY(&pgrp->pg_members)); 1050 1051 pt->pt_pgrp = NULL; 1052 1053 if (!P_VALID(pt->pt_proc)) { 1054 /* Orphaned pgrp, put slot onto free list. */ 1055 KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0); 1056 pg_id &= pid_tbl_mask; 1057 pt = &pid_table[last_free_pt]; 1058 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id); 1059 KASSERT(pt->pt_pid == 0); 1060 last_free_pt = pg_id; 1061 pid_alloc_cnt--; 1062 } 1063 return pgrp; 1064 } 1065 1066 /* 1067 * pg_delete: delete and free a process group. 1068 * => must be called with the proc_lock held, which will be released. 1069 */ 1070 static void 1071 pg_delete(pid_t pg_id) 1072 { 1073 struct pgrp *pg; 1074 struct tty *ttyp; 1075 struct session *ss; 1076 1077 KASSERT(mutex_owned(proc_lock)); 1078 1079 pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp; 1080 if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) { 1081 mutex_exit(proc_lock); 1082 return; 1083 } 1084 1085 ss = pg->pg_session; 1086 1087 /* Remove reference (if any) from tty to this process group */ 1088 mutex_spin_enter(&tty_lock); 1089 ttyp = ss->s_ttyp; 1090 if (ttyp != NULL && ttyp->t_pgrp == pg) { 1091 ttyp->t_pgrp = NULL; 1092 KASSERT(ttyp->t_session == ss); 1093 } 1094 mutex_spin_exit(&tty_lock); 1095 1096 /* 1097 * The leading process group in a session is freed by proc_sessrele(), 1098 * if last reference. Note: proc_sessrele() releases proc_lock. 1099 */ 1100 pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL; 1101 proc_sessrele(ss); 1102 1103 if (pg != NULL) { 1104 /* Free it, if was not done by proc_sessrele(). */ 1105 kmem_free(pg, sizeof(struct pgrp)); 1106 } 1107 } 1108 1109 /* 1110 * Adjust pgrp jobc counters when specified process changes process group. 1111 * We count the number of processes in each process group that "qualify" 1112 * the group for terminal job control (those with a parent in a different 1113 * process group of the same session). If that count reaches zero, the 1114 * process group becomes orphaned. Check both the specified process' 1115 * process group and that of its children. 1116 * entering == 0 => p is leaving specified group. 1117 * entering == 1 => p is entering specified group. 1118 * 1119 * Call with proc_lock held. 1120 */ 1121 void 1122 fixjobc(struct proc *p, struct pgrp *pgrp, int entering) 1123 { 1124 struct pgrp *hispgrp; 1125 struct session *mysession = pgrp->pg_session; 1126 struct proc *child; 1127 1128 KASSERT(mutex_owned(proc_lock)); 1129 1130 /* 1131 * Check p's parent to see whether p qualifies its own process 1132 * group; if so, adjust count for p's process group. 1133 */ 1134 hispgrp = p->p_pptr->p_pgrp; 1135 if (hispgrp != pgrp && hispgrp->pg_session == mysession) { 1136 if (entering) { 1137 pgrp->pg_jobc++; 1138 p->p_lflag &= ~PL_ORPHANPG; 1139 } else if (--pgrp->pg_jobc == 0) 1140 orphanpg(pgrp); 1141 } 1142 1143 /* 1144 * Check this process' children to see whether they qualify 1145 * their process groups; if so, adjust counts for children's 1146 * process groups. 1147 */ 1148 LIST_FOREACH(child, &p->p_children, p_sibling) { 1149 hispgrp = child->p_pgrp; 1150 if (hispgrp != pgrp && hispgrp->pg_session == mysession && 1151 !P_ZOMBIE(child)) { 1152 if (entering) { 1153 child->p_lflag &= ~PL_ORPHANPG; 1154 hispgrp->pg_jobc++; 1155 } else if (--hispgrp->pg_jobc == 0) 1156 orphanpg(hispgrp); 1157 } 1158 } 1159 } 1160 1161 /* 1162 * A process group has become orphaned; 1163 * if there are any stopped processes in the group, 1164 * hang-up all process in that group. 1165 * 1166 * Call with proc_lock held. 1167 */ 1168 static void 1169 orphanpg(struct pgrp *pg) 1170 { 1171 struct proc *p; 1172 1173 KASSERT(mutex_owned(proc_lock)); 1174 1175 LIST_FOREACH(p, &pg->pg_members, p_pglist) { 1176 if (p->p_stat == SSTOP) { 1177 p->p_lflag |= PL_ORPHANPG; 1178 psignal(p, SIGHUP); 1179 psignal(p, SIGCONT); 1180 } 1181 } 1182 } 1183 1184 #ifdef DDB 1185 #include <ddb/db_output.h> 1186 void pidtbl_dump(void); 1187 void 1188 pidtbl_dump(void) 1189 { 1190 struct pid_table *pt; 1191 struct proc *p; 1192 struct pgrp *pgrp; 1193 int id; 1194 1195 db_printf("pid table %p size %x, next %x, last %x\n", 1196 pid_table, pid_tbl_mask+1, 1197 next_free_pt, last_free_pt); 1198 for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) { 1199 p = pt->pt_proc; 1200 if (!P_VALID(p) && !pt->pt_pgrp) 1201 continue; 1202 db_printf(" id %x: ", id); 1203 if (P_VALID(p)) 1204 db_printf("slotpid %d proc %p id %d (0x%x) %s\n", 1205 pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm); 1206 else 1207 db_printf("next %x use %x\n", 1208 P_NEXT(p) & pid_tbl_mask, 1209 P_NEXT(p) & ~pid_tbl_mask); 1210 if ((pgrp = pt->pt_pgrp)) { 1211 db_printf("\tsession %p, sid %d, count %d, login %s\n", 1212 pgrp->pg_session, pgrp->pg_session->s_sid, 1213 pgrp->pg_session->s_count, 1214 pgrp->pg_session->s_login); 1215 db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n", 1216 pgrp, pgrp->pg_id, pgrp->pg_jobc, 1217 LIST_FIRST(&pgrp->pg_members)); 1218 LIST_FOREACH(p, &pgrp->pg_members, p_pglist) { 1219 db_printf("\t\tpid %d addr %p pgrp %p %s\n", 1220 p->p_pid, p, p->p_pgrp, p->p_comm); 1221 } 1222 } 1223 } 1224 } 1225 #endif /* DDB */ 1226 1227 #ifdef KSTACK_CHECK_MAGIC 1228 1229 #define KSTACK_MAGIC 0xdeadbeaf 1230 1231 /* XXX should be per process basis? */ 1232 static int kstackleftmin = KSTACK_SIZE; 1233 static int kstackleftthres = KSTACK_SIZE / 8; 1234 1235 void 1236 kstack_setup_magic(const struct lwp *l) 1237 { 1238 uint32_t *ip; 1239 uint32_t const *end; 1240 1241 KASSERT(l != NULL); 1242 KASSERT(l != &lwp0); 1243 1244 /* 1245 * fill all the stack with magic number 1246 * so that later modification on it can be detected. 1247 */ 1248 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1249 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1250 for (; ip < end; ip++) { 1251 *ip = KSTACK_MAGIC; 1252 } 1253 } 1254 1255 void 1256 kstack_check_magic(const struct lwp *l) 1257 { 1258 uint32_t const *ip, *end; 1259 int stackleft; 1260 1261 KASSERT(l != NULL); 1262 1263 /* don't check proc0 */ /*XXX*/ 1264 if (l == &lwp0) 1265 return; 1266 1267 #ifdef __MACHINE_STACK_GROWS_UP 1268 /* stack grows upwards (eg. hppa) */ 1269 ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1270 end = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1271 for (ip--; ip >= end; ip--) 1272 if (*ip != KSTACK_MAGIC) 1273 break; 1274 1275 stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip; 1276 #else /* __MACHINE_STACK_GROWS_UP */ 1277 /* stack grows downwards (eg. i386) */ 1278 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1279 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1280 for (; ip < end; ip++) 1281 if (*ip != KSTACK_MAGIC) 1282 break; 1283 1284 stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l); 1285 #endif /* __MACHINE_STACK_GROWS_UP */ 1286 1287 if (kstackleftmin > stackleft) { 1288 kstackleftmin = stackleft; 1289 if (stackleft < kstackleftthres) 1290 printf("warning: kernel stack left %d bytes" 1291 "(pid %u:lid %u)\n", stackleft, 1292 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1293 } 1294 1295 if (stackleft <= 0) { 1296 panic("magic on the top of kernel stack changed for " 1297 "pid %u, lid %u: maybe kernel stack overflow", 1298 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1299 } 1300 } 1301 #endif /* KSTACK_CHECK_MAGIC */ 1302 1303 int 1304 proclist_foreach_call(struct proclist *list, 1305 int (*callback)(struct proc *, void *arg), void *arg) 1306 { 1307 struct proc marker; 1308 struct proc *p; 1309 int ret = 0; 1310 1311 marker.p_flag = PK_MARKER; 1312 mutex_enter(proc_lock); 1313 for (p = LIST_FIRST(list); ret == 0 && p != NULL;) { 1314 if (p->p_flag & PK_MARKER) { 1315 p = LIST_NEXT(p, p_list); 1316 continue; 1317 } 1318 LIST_INSERT_AFTER(p, &marker, p_list); 1319 ret = (*callback)(p, arg); 1320 KASSERT(mutex_owned(proc_lock)); 1321 p = LIST_NEXT(&marker, p_list); 1322 LIST_REMOVE(&marker, p_list); 1323 } 1324 mutex_exit(proc_lock); 1325 1326 return ret; 1327 } 1328 1329 int 1330 proc_vmspace_getref(struct proc *p, struct vmspace **vm) 1331 { 1332 1333 /* XXXCDC: how should locking work here? */ 1334 1335 /* curproc exception is for coredump. */ 1336 1337 if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) || 1338 (p->p_vmspace->vm_refcnt < 1)) { /* XXX */ 1339 return EFAULT; 1340 } 1341 1342 uvmspace_addref(p->p_vmspace); 1343 *vm = p->p_vmspace; 1344 1345 return 0; 1346 } 1347 1348 /* 1349 * Acquire a write lock on the process credential. 1350 */ 1351 void 1352 proc_crmod_enter(void) 1353 { 1354 struct lwp *l = curlwp; 1355 struct proc *p = l->l_proc; 1356 kauth_cred_t oc; 1357 1358 /* Reset what needs to be reset in plimit. */ 1359 if (p->p_limit->pl_corename != defcorename) { 1360 lim_setcorename(p, defcorename, 0); 1361 } 1362 1363 mutex_enter(p->p_lock); 1364 1365 /* Ensure the LWP cached credentials are up to date. */ 1366 if ((oc = l->l_cred) != p->p_cred) { 1367 kauth_cred_hold(p->p_cred); 1368 l->l_cred = p->p_cred; 1369 kauth_cred_free(oc); 1370 } 1371 } 1372 1373 /* 1374 * Set in a new process credential, and drop the write lock. The credential 1375 * must have a reference already. Optionally, free a no-longer required 1376 * credential. The scheduler also needs to inspect p_cred, so we also 1377 * briefly acquire the sched state mutex. 1378 */ 1379 void 1380 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid) 1381 { 1382 struct lwp *l = curlwp, *l2; 1383 struct proc *p = l->l_proc; 1384 kauth_cred_t oc; 1385 1386 KASSERT(mutex_owned(p->p_lock)); 1387 1388 /* Is there a new credential to set in? */ 1389 if (scred != NULL) { 1390 p->p_cred = scred; 1391 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 1392 if (l2 != l) 1393 l2->l_prflag |= LPR_CRMOD; 1394 } 1395 1396 /* Ensure the LWP cached credentials are up to date. */ 1397 if ((oc = l->l_cred) != scred) { 1398 kauth_cred_hold(scred); 1399 l->l_cred = scred; 1400 } 1401 } else 1402 oc = NULL; /* XXXgcc */ 1403 1404 if (sugid) { 1405 /* 1406 * Mark process as having changed credentials, stops 1407 * tracing etc. 1408 */ 1409 p->p_flag |= PK_SUGID; 1410 } 1411 1412 mutex_exit(p->p_lock); 1413 1414 /* If there is a credential to be released, free it now. */ 1415 if (fcred != NULL) { 1416 KASSERT(scred != NULL); 1417 kauth_cred_free(fcred); 1418 if (oc != scred) 1419 kauth_cred_free(oc); 1420 } 1421 } 1422 1423 /* 1424 * proc_specific_key_create -- 1425 * Create a key for subsystem proc-specific data. 1426 */ 1427 int 1428 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor) 1429 { 1430 1431 return (specificdata_key_create(proc_specificdata_domain, keyp, dtor)); 1432 } 1433 1434 /* 1435 * proc_specific_key_delete -- 1436 * Delete a key for subsystem proc-specific data. 1437 */ 1438 void 1439 proc_specific_key_delete(specificdata_key_t key) 1440 { 1441 1442 specificdata_key_delete(proc_specificdata_domain, key); 1443 } 1444 1445 /* 1446 * proc_initspecific -- 1447 * Initialize a proc's specificdata container. 1448 */ 1449 void 1450 proc_initspecific(struct proc *p) 1451 { 1452 int error; 1453 1454 error = specificdata_init(proc_specificdata_domain, &p->p_specdataref); 1455 KASSERT(error == 0); 1456 } 1457 1458 /* 1459 * proc_finispecific -- 1460 * Finalize a proc's specificdata container. 1461 */ 1462 void 1463 proc_finispecific(struct proc *p) 1464 { 1465 1466 specificdata_fini(proc_specificdata_domain, &p->p_specdataref); 1467 } 1468 1469 /* 1470 * proc_getspecific -- 1471 * Return proc-specific data corresponding to the specified key. 1472 */ 1473 void * 1474 proc_getspecific(struct proc *p, specificdata_key_t key) 1475 { 1476 1477 return (specificdata_getspecific(proc_specificdata_domain, 1478 &p->p_specdataref, key)); 1479 } 1480 1481 /* 1482 * proc_setspecific -- 1483 * Set proc-specific data corresponding to the specified key. 1484 */ 1485 void 1486 proc_setspecific(struct proc *p, specificdata_key_t key, void *data) 1487 { 1488 1489 specificdata_setspecific(proc_specificdata_domain, 1490 &p->p_specdataref, key, data); 1491 } 1492 1493 int 1494 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target) 1495 { 1496 int r = 0; 1497 1498 if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) || 1499 kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) { 1500 /* 1501 * suid proc of ours or proc not ours 1502 */ 1503 r = EPERM; 1504 } else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) { 1505 /* 1506 * sgid proc has sgid back to us temporarily 1507 */ 1508 r = EPERM; 1509 } else { 1510 /* 1511 * our rgid must be in target's group list (ie, 1512 * sub-processes started by a sgid process) 1513 */ 1514 int ismember = 0; 1515 1516 if (kauth_cred_ismember_gid(cred, 1517 kauth_cred_getgid(target), &ismember) != 0 || 1518 !ismember) 1519 r = EPERM; 1520 } 1521 1522 return (r); 1523 } 1524 1525 /* 1526 * sysctl stuff 1527 */ 1528 1529 #define KERN_PROCSLOP (5 * sizeof(struct kinfo_proc)) 1530 1531 static const u_int sysctl_flagmap[] = { 1532 PK_ADVLOCK, P_ADVLOCK, 1533 PK_EXEC, P_EXEC, 1534 PK_NOCLDWAIT, P_NOCLDWAIT, 1535 PK_32, P_32, 1536 PK_CLDSIGIGN, P_CLDSIGIGN, 1537 PK_SUGID, P_SUGID, 1538 0 1539 }; 1540 1541 static const u_int sysctl_sflagmap[] = { 1542 PS_NOCLDSTOP, P_NOCLDSTOP, 1543 PS_WEXIT, P_WEXIT, 1544 PS_STOPFORK, P_STOPFORK, 1545 PS_STOPEXEC, P_STOPEXEC, 1546 PS_STOPEXIT, P_STOPEXIT, 1547 0 1548 }; 1549 1550 static const u_int sysctl_slflagmap[] = { 1551 PSL_TRACED, P_TRACED, 1552 PSL_FSTRACE, P_FSTRACE, 1553 PSL_CHTRACED, P_CHTRACED, 1554 PSL_SYSCALL, P_SYSCALL, 1555 0 1556 }; 1557 1558 static const u_int sysctl_lflagmap[] = { 1559 PL_CONTROLT, P_CONTROLT, 1560 PL_PPWAIT, P_PPWAIT, 1561 0 1562 }; 1563 1564 static const u_int sysctl_stflagmap[] = { 1565 PST_PROFIL, P_PROFIL, 1566 0 1567 1568 }; 1569 1570 /* used by kern_lwp also */ 1571 const u_int sysctl_lwpflagmap[] = { 1572 LW_SINTR, L_SINTR, 1573 LW_SYSTEM, L_SYSTEM, 1574 0 1575 }; 1576 1577 /* 1578 * Find the most ``active'' lwp of a process and return it for ps display 1579 * purposes 1580 */ 1581 static struct lwp * 1582 proc_active_lwp(struct proc *p) 1583 { 1584 static const int ostat[] = { 1585 0, 1586 2, /* LSIDL */ 1587 6, /* LSRUN */ 1588 5, /* LSSLEEP */ 1589 4, /* LSSTOP */ 1590 0, /* LSZOMB */ 1591 1, /* LSDEAD */ 1592 7, /* LSONPROC */ 1593 3 /* LSSUSPENDED */ 1594 }; 1595 1596 struct lwp *l, *lp = NULL; 1597 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 1598 KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat)); 1599 if (lp == NULL || 1600 ostat[l->l_stat] > ostat[lp->l_stat] || 1601 (ostat[l->l_stat] == ostat[lp->l_stat] && 1602 l->l_cpticks > lp->l_cpticks)) { 1603 lp = l; 1604 continue; 1605 } 1606 } 1607 return lp; 1608 } 1609 1610 static int 1611 sysctl_doeproc(SYSCTLFN_ARGS) 1612 { 1613 union { 1614 struct kinfo_proc kproc; 1615 struct kinfo_proc2 kproc2; 1616 } *kbuf; 1617 struct proc *p, *next, *marker; 1618 char *where, *dp; 1619 int type, op, arg, error; 1620 u_int elem_size, kelem_size, elem_count; 1621 size_t buflen, needed; 1622 bool match, zombie, mmmbrains; 1623 1624 if (namelen == 1 && name[0] == CTL_QUERY) 1625 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1626 1627 dp = where = oldp; 1628 buflen = where != NULL ? *oldlenp : 0; 1629 error = 0; 1630 needed = 0; 1631 type = rnode->sysctl_num; 1632 1633 if (type == KERN_PROC) { 1634 if (namelen != 2 && !(namelen == 1 && name[0] == KERN_PROC_ALL)) 1635 return (EINVAL); 1636 op = name[0]; 1637 if (op != KERN_PROC_ALL) 1638 arg = name[1]; 1639 else 1640 arg = 0; /* Quell compiler warning */ 1641 elem_count = 0; /* Ditto */ 1642 kelem_size = elem_size = sizeof(kbuf->kproc); 1643 } else { 1644 if (namelen != 4) 1645 return (EINVAL); 1646 op = name[0]; 1647 arg = name[1]; 1648 elem_size = name[2]; 1649 elem_count = name[3]; 1650 kelem_size = sizeof(kbuf->kproc2); 1651 } 1652 1653 sysctl_unlock(); 1654 1655 kbuf = kmem_alloc(sizeof(*kbuf), KM_SLEEP); 1656 marker = kmem_alloc(sizeof(*marker), KM_SLEEP); 1657 marker->p_flag = PK_MARKER; 1658 1659 mutex_enter(proc_lock); 1660 mmmbrains = false; 1661 for (p = LIST_FIRST(&allproc);; p = next) { 1662 if (p == NULL) { 1663 if (!mmmbrains) { 1664 p = LIST_FIRST(&zombproc); 1665 mmmbrains = true; 1666 } 1667 if (p == NULL) 1668 break; 1669 } 1670 next = LIST_NEXT(p, p_list); 1671 if ((p->p_flag & PK_MARKER) != 0) 1672 continue; 1673 1674 /* 1675 * Skip embryonic processes. 1676 */ 1677 if (p->p_stat == SIDL) 1678 continue; 1679 1680 mutex_enter(p->p_lock); 1681 error = kauth_authorize_process(l->l_cred, 1682 KAUTH_PROCESS_CANSEE, p, 1683 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL); 1684 if (error != 0) { 1685 mutex_exit(p->p_lock); 1686 continue; 1687 } 1688 1689 /* 1690 * TODO - make more efficient (see notes below). 1691 * do by session. 1692 */ 1693 switch (op) { 1694 case KERN_PROC_PID: 1695 /* could do this with just a lookup */ 1696 match = (p->p_pid == (pid_t)arg); 1697 break; 1698 1699 case KERN_PROC_PGRP: 1700 /* could do this by traversing pgrp */ 1701 match = (p->p_pgrp->pg_id == (pid_t)arg); 1702 break; 1703 1704 case KERN_PROC_SESSION: 1705 match = (p->p_session->s_sid == (pid_t)arg); 1706 break; 1707 1708 case KERN_PROC_TTY: 1709 match = true; 1710 if (arg == (int) KERN_PROC_TTY_REVOKE) { 1711 if ((p->p_lflag & PL_CONTROLT) == 0 || 1712 p->p_session->s_ttyp == NULL || 1713 p->p_session->s_ttyvp != NULL) { 1714 match = false; 1715 } 1716 } else if ((p->p_lflag & PL_CONTROLT) == 0 || 1717 p->p_session->s_ttyp == NULL) { 1718 if ((dev_t)arg != KERN_PROC_TTY_NODEV) { 1719 match = false; 1720 } 1721 } else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) { 1722 match = false; 1723 } 1724 break; 1725 1726 case KERN_PROC_UID: 1727 match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg); 1728 break; 1729 1730 case KERN_PROC_RUID: 1731 match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg); 1732 break; 1733 1734 case KERN_PROC_GID: 1735 match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg); 1736 break; 1737 1738 case KERN_PROC_RGID: 1739 match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg); 1740 break; 1741 1742 case KERN_PROC_ALL: 1743 match = true; 1744 /* allow everything */ 1745 break; 1746 1747 default: 1748 error = EINVAL; 1749 mutex_exit(p->p_lock); 1750 goto cleanup; 1751 } 1752 if (!match) { 1753 mutex_exit(p->p_lock); 1754 continue; 1755 } 1756 1757 /* 1758 * Grab a hold on the process. 1759 */ 1760 if (mmmbrains) { 1761 zombie = true; 1762 } else { 1763 zombie = !rw_tryenter(&p->p_reflock, RW_READER); 1764 } 1765 if (zombie) { 1766 LIST_INSERT_AFTER(p, marker, p_list); 1767 } 1768 1769 if (buflen >= elem_size && 1770 (type == KERN_PROC || elem_count > 0)) { 1771 if (type == KERN_PROC) { 1772 kbuf->kproc.kp_proc = *p; 1773 fill_eproc(p, &kbuf->kproc.kp_eproc, zombie); 1774 } else { 1775 fill_kproc2(p, &kbuf->kproc2, zombie); 1776 elem_count--; 1777 } 1778 mutex_exit(p->p_lock); 1779 mutex_exit(proc_lock); 1780 /* 1781 * Copy out elem_size, but not larger than kelem_size 1782 */ 1783 error = sysctl_copyout(l, kbuf, dp, 1784 min(kelem_size, elem_size)); 1785 mutex_enter(proc_lock); 1786 if (error) { 1787 goto bah; 1788 } 1789 dp += elem_size; 1790 buflen -= elem_size; 1791 } else { 1792 mutex_exit(p->p_lock); 1793 } 1794 needed += elem_size; 1795 1796 /* 1797 * Release reference to process. 1798 */ 1799 if (zombie) { 1800 next = LIST_NEXT(marker, p_list); 1801 LIST_REMOVE(marker, p_list); 1802 } else { 1803 rw_exit(&p->p_reflock); 1804 next = LIST_NEXT(p, p_list); 1805 } 1806 } 1807 mutex_exit(proc_lock); 1808 1809 if (where != NULL) { 1810 *oldlenp = dp - where; 1811 if (needed > *oldlenp) { 1812 error = ENOMEM; 1813 goto out; 1814 } 1815 } else { 1816 needed += KERN_PROCSLOP; 1817 *oldlenp = needed; 1818 } 1819 if (kbuf) 1820 kmem_free(kbuf, sizeof(*kbuf)); 1821 if (marker) 1822 kmem_free(marker, sizeof(*marker)); 1823 sysctl_relock(); 1824 return 0; 1825 bah: 1826 if (zombie) 1827 LIST_REMOVE(marker, p_list); 1828 else 1829 rw_exit(&p->p_reflock); 1830 cleanup: 1831 mutex_exit(proc_lock); 1832 out: 1833 if (kbuf) 1834 kmem_free(kbuf, sizeof(*kbuf)); 1835 if (marker) 1836 kmem_free(marker, sizeof(*marker)); 1837 sysctl_relock(); 1838 return error; 1839 } 1840 1841 int 1842 copyin_psstrings(struct proc *p, struct ps_strings *arginfo) 1843 { 1844 1845 #ifdef COMPAT_NETBSD32 1846 if (p->p_flag & PK_32) { 1847 struct ps_strings32 arginfo32; 1848 1849 int error = copyin_proc(p, (void *)p->p_psstrp, &arginfo32, 1850 sizeof(arginfo32)); 1851 if (error) 1852 return error; 1853 arginfo->ps_argvstr = (void *)(uintptr_t)arginfo32.ps_argvstr; 1854 arginfo->ps_nargvstr = arginfo32.ps_nargvstr; 1855 arginfo->ps_envstr = (void *)(uintptr_t)arginfo32.ps_envstr; 1856 arginfo->ps_nenvstr = arginfo32.ps_nenvstr; 1857 return 0; 1858 } 1859 #endif 1860 return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo)); 1861 } 1862 1863 static int 1864 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len) 1865 { 1866 void **cookie = cookie_; 1867 struct lwp *l = cookie[0]; 1868 char *dst = cookie[1]; 1869 1870 return sysctl_copyout(l, src, dst + off, len); 1871 } 1872 1873 /* 1874 * sysctl helper routine for kern.proc_args pseudo-subtree. 1875 */ 1876 static int 1877 sysctl_kern_proc_args(SYSCTLFN_ARGS) 1878 { 1879 struct ps_strings pss; 1880 struct proc *p; 1881 pid_t pid; 1882 int type, error; 1883 void *cookie[2]; 1884 1885 if (namelen == 1 && name[0] == CTL_QUERY) 1886 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1887 1888 if (newp != NULL || namelen != 2) 1889 return (EINVAL); 1890 pid = name[0]; 1891 type = name[1]; 1892 1893 switch (type) { 1894 case KERN_PROC_ARGV: 1895 case KERN_PROC_NARGV: 1896 case KERN_PROC_ENV: 1897 case KERN_PROC_NENV: 1898 /* ok */ 1899 break; 1900 default: 1901 return (EINVAL); 1902 } 1903 1904 sysctl_unlock(); 1905 1906 /* check pid */ 1907 mutex_enter(proc_lock); 1908 if ((p = proc_find(pid)) == NULL) { 1909 error = EINVAL; 1910 goto out_locked; 1911 } 1912 mutex_enter(p->p_lock); 1913 1914 /* Check permission. */ 1915 if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV) 1916 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1917 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL); 1918 else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV) 1919 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1920 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL); 1921 else 1922 error = EINVAL; /* XXXGCC */ 1923 if (error) { 1924 mutex_exit(p->p_lock); 1925 goto out_locked; 1926 } 1927 1928 if (oldp == NULL) { 1929 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) 1930 *oldlenp = sizeof (int); 1931 else 1932 *oldlenp = ARG_MAX; /* XXX XXX XXX */ 1933 error = 0; 1934 mutex_exit(p->p_lock); 1935 goto out_locked; 1936 } 1937 1938 /* 1939 * Zombies don't have a stack, so we can't read their psstrings. 1940 * System processes also don't have a user stack. 1941 */ 1942 if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) { 1943 error = EINVAL; 1944 mutex_exit(p->p_lock); 1945 goto out_locked; 1946 } 1947 1948 error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY; 1949 mutex_exit(p->p_lock); 1950 if (error) { 1951 goto out_locked; 1952 } 1953 mutex_exit(proc_lock); 1954 1955 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) { 1956 int value; 1957 if ((error = copyin_psstrings(p, &pss)) == 0) { 1958 if (type == KERN_PROC_NARGV) 1959 value = pss.ps_nargvstr; 1960 else 1961 value = pss.ps_nenvstr; 1962 error = sysctl_copyout(l, &value, oldp, sizeof(value)); 1963 *oldlenp = sizeof(value); 1964 } 1965 } else { 1966 cookie[0] = l; 1967 cookie[1] = oldp; 1968 error = copy_procargs(p, type, oldlenp, 1969 copy_procargs_sysctl_cb, cookie); 1970 } 1971 rw_exit(&p->p_reflock); 1972 sysctl_relock(); 1973 return error; 1974 1975 out_locked: 1976 mutex_exit(proc_lock); 1977 sysctl_relock(); 1978 return error; 1979 } 1980 1981 int 1982 copy_procargs(struct proc *p, int oid, size_t *limit, 1983 int (*cb)(void *, const void *, size_t, size_t), void *cookie) 1984 { 1985 struct ps_strings pss; 1986 size_t len, i, loaded, entry_len; 1987 struct uio auio; 1988 struct iovec aiov; 1989 int error, argvlen; 1990 char *arg; 1991 char **argv; 1992 vaddr_t user_argv; 1993 struct vmspace *vmspace; 1994 1995 /* 1996 * Allocate a temporary buffer to hold the argument vector and 1997 * the arguments themselve. 1998 */ 1999 arg = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2000 argv = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2001 2002 /* 2003 * Lock the process down in memory. 2004 */ 2005 vmspace = p->p_vmspace; 2006 uvmspace_addref(vmspace); 2007 2008 /* 2009 * Read in the ps_strings structure. 2010 */ 2011 if ((error = copyin_psstrings(p, &pss)) != 0) 2012 goto done; 2013 2014 /* 2015 * Now read the address of the argument vector. 2016 */ 2017 switch (oid) { 2018 case KERN_PROC_ARGV: 2019 user_argv = (uintptr_t)pss.ps_argvstr; 2020 argvlen = pss.ps_nargvstr; 2021 break; 2022 case KERN_PROC_ENV: 2023 user_argv = (uintptr_t)pss.ps_envstr; 2024 argvlen = pss.ps_nenvstr; 2025 break; 2026 default: 2027 error = EINVAL; 2028 goto done; 2029 } 2030 2031 if (argvlen < 0) { 2032 error = EIO; 2033 goto done; 2034 } 2035 2036 #ifdef COMPAT_NETBSD32 2037 if (p->p_flag & PK_32) 2038 entry_len = sizeof(netbsd32_charp); 2039 else 2040 #endif 2041 entry_len = sizeof(char *); 2042 2043 /* 2044 * Now copy each string. 2045 */ 2046 len = 0; /* bytes written to user buffer */ 2047 loaded = 0; /* bytes from argv already processed */ 2048 i = 0; /* To make compiler happy */ 2049 2050 for (; argvlen; --argvlen) { 2051 int finished = 0; 2052 vaddr_t base; 2053 size_t xlen; 2054 int j; 2055 2056 if (loaded == 0) { 2057 size_t rem = entry_len * argvlen; 2058 loaded = MIN(rem, PAGE_SIZE); 2059 error = copyin_vmspace(vmspace, 2060 (const void *)user_argv, argv, loaded); 2061 if (error) 2062 break; 2063 user_argv += loaded; 2064 i = 0; 2065 } 2066 2067 #ifdef COMPAT_NETBSD32 2068 if (p->p_flag & PK_32) { 2069 netbsd32_charp *argv32; 2070 2071 argv32 = (netbsd32_charp *)argv; 2072 base = (vaddr_t)NETBSD32PTR64(argv32[i++]); 2073 } else 2074 #endif 2075 base = (vaddr_t)argv[i++]; 2076 loaded -= entry_len; 2077 2078 /* 2079 * The program has messed around with its arguments, 2080 * possibly deleting some, and replacing them with 2081 * NULL's. Treat this as the last argument and not 2082 * a failure. 2083 */ 2084 if (base == 0) 2085 break; 2086 2087 while (!finished) { 2088 xlen = PAGE_SIZE - (base & PAGE_MASK); 2089 2090 aiov.iov_base = arg; 2091 aiov.iov_len = PAGE_SIZE; 2092 auio.uio_iov = &aiov; 2093 auio.uio_iovcnt = 1; 2094 auio.uio_offset = base; 2095 auio.uio_resid = xlen; 2096 auio.uio_rw = UIO_READ; 2097 UIO_SETUP_SYSSPACE(&auio); 2098 error = uvm_io(&vmspace->vm_map, &auio); 2099 if (error) 2100 goto done; 2101 2102 /* Look for the end of the string */ 2103 for (j = 0; j < xlen; j++) { 2104 if (arg[j] == '\0') { 2105 xlen = j + 1; 2106 finished = 1; 2107 break; 2108 } 2109 } 2110 2111 /* Check for user buffer overflow */ 2112 if (len + xlen > *limit) { 2113 finished = 1; 2114 if (len > *limit) 2115 xlen = 0; 2116 else 2117 xlen = *limit - len; 2118 } 2119 2120 /* Copyout the page */ 2121 error = (*cb)(cookie, arg, len, xlen); 2122 if (error) 2123 goto done; 2124 2125 len += xlen; 2126 base += xlen; 2127 } 2128 } 2129 *limit = len; 2130 2131 done: 2132 kmem_free(argv, PAGE_SIZE); 2133 kmem_free(arg, PAGE_SIZE); 2134 uvmspace_free(vmspace); 2135 return error; 2136 } 2137 2138 /* 2139 * Fill in an eproc structure for the specified process. 2140 */ 2141 void 2142 fill_eproc(struct proc *p, struct eproc *ep, bool zombie) 2143 { 2144 struct tty *tp; 2145 struct lwp *l; 2146 2147 KASSERT(mutex_owned(proc_lock)); 2148 KASSERT(mutex_owned(p->p_lock)); 2149 2150 memset(ep, 0, sizeof(*ep)); 2151 2152 ep->e_paddr = p; 2153 ep->e_sess = p->p_session; 2154 if (p->p_cred) { 2155 kauth_cred_topcred(p->p_cred, &ep->e_pcred); 2156 kauth_cred_toucred(p->p_cred, &ep->e_ucred); 2157 } 2158 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2159 struct vmspace *vm = p->p_vmspace; 2160 2161 ep->e_vm.vm_rssize = vm_resident_count(vm); 2162 ep->e_vm.vm_tsize = vm->vm_tsize; 2163 ep->e_vm.vm_dsize = vm->vm_dsize; 2164 ep->e_vm.vm_ssize = vm->vm_ssize; 2165 ep->e_vm.vm_map.size = vm->vm_map.size; 2166 2167 /* Pick the primary (first) LWP */ 2168 l = proc_active_lwp(p); 2169 KASSERT(l != NULL); 2170 lwp_lock(l); 2171 if (l->l_wchan) 2172 strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN); 2173 lwp_unlock(l); 2174 } 2175 if (p->p_pptr) 2176 ep->e_ppid = p->p_pptr->p_pid; 2177 if (p->p_pgrp && p->p_session) { 2178 ep->e_pgid = p->p_pgrp->pg_id; 2179 ep->e_jobc = p->p_pgrp->pg_jobc; 2180 ep->e_sid = p->p_session->s_sid; 2181 if ((p->p_lflag & PL_CONTROLT) && 2182 (tp = ep->e_sess->s_ttyp)) { 2183 ep->e_tdev = tp->t_dev; 2184 ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2185 ep->e_tsess = tp->t_session; 2186 } else 2187 ep->e_tdev = (uint32_t)NODEV; 2188 ep->e_flag = ep->e_sess->s_ttyvp ? EPROC_CTTY : 0; 2189 if (SESS_LEADER(p)) 2190 ep->e_flag |= EPROC_SLEADER; 2191 strncpy(ep->e_login, ep->e_sess->s_login, MAXLOGNAME); 2192 } 2193 ep->e_xsize = ep->e_xrssize = 0; 2194 ep->e_xccount = ep->e_xswrss = 0; 2195 } 2196 2197 /* 2198 * Fill in a kinfo_proc2 structure for the specified process. 2199 */ 2200 static void 2201 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie) 2202 { 2203 struct tty *tp; 2204 struct lwp *l, *l2; 2205 struct timeval ut, st, rt; 2206 sigset_t ss1, ss2; 2207 struct rusage ru; 2208 struct vmspace *vm; 2209 2210 KASSERT(mutex_owned(proc_lock)); 2211 KASSERT(mutex_owned(p->p_lock)); 2212 2213 sigemptyset(&ss1); 2214 sigemptyset(&ss2); 2215 memset(ki, 0, sizeof(*ki)); 2216 2217 ki->p_paddr = PTRTOUINT64(p); 2218 ki->p_fd = PTRTOUINT64(p->p_fd); 2219 ki->p_cwdi = PTRTOUINT64(p->p_cwdi); 2220 ki->p_stats = PTRTOUINT64(p->p_stats); 2221 ki->p_limit = PTRTOUINT64(p->p_limit); 2222 ki->p_vmspace = PTRTOUINT64(p->p_vmspace); 2223 ki->p_sigacts = PTRTOUINT64(p->p_sigacts); 2224 ki->p_sess = PTRTOUINT64(p->p_session); 2225 ki->p_tsess = 0; /* may be changed if controlling tty below */ 2226 ki->p_ru = PTRTOUINT64(&p->p_stats->p_ru); 2227 ki->p_eflag = 0; 2228 ki->p_exitsig = p->p_exitsig; 2229 ki->p_flag = L_INMEM; /* Process never swapped out */ 2230 ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag); 2231 ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag); 2232 ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag); 2233 ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag); 2234 ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag); 2235 ki->p_pid = p->p_pid; 2236 if (p->p_pptr) 2237 ki->p_ppid = p->p_pptr->p_pid; 2238 else 2239 ki->p_ppid = 0; 2240 ki->p_uid = kauth_cred_geteuid(p->p_cred); 2241 ki->p_ruid = kauth_cred_getuid(p->p_cred); 2242 ki->p_gid = kauth_cred_getegid(p->p_cred); 2243 ki->p_rgid = kauth_cred_getgid(p->p_cred); 2244 ki->p_svuid = kauth_cred_getsvuid(p->p_cred); 2245 ki->p_svgid = kauth_cred_getsvgid(p->p_cred); 2246 ki->p_ngroups = kauth_cred_ngroups(p->p_cred); 2247 kauth_cred_getgroups(p->p_cred, ki->p_groups, 2248 min(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])), 2249 UIO_SYSSPACE); 2250 2251 ki->p_uticks = p->p_uticks; 2252 ki->p_sticks = p->p_sticks; 2253 ki->p_iticks = p->p_iticks; 2254 ki->p_tpgid = NO_PGID; /* may be changed if controlling tty below */ 2255 ki->p_tracep = PTRTOUINT64(p->p_tracep); 2256 ki->p_traceflag = p->p_traceflag; 2257 2258 memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t)); 2259 memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t)); 2260 2261 ki->p_cpticks = 0; 2262 ki->p_pctcpu = p->p_pctcpu; 2263 ki->p_estcpu = 0; 2264 ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */ 2265 ki->p_realstat = p->p_stat; 2266 ki->p_nice = p->p_nice; 2267 ki->p_xstat = p->p_xstat; 2268 ki->p_acflag = p->p_acflag; 2269 2270 strncpy(ki->p_comm, p->p_comm, 2271 min(sizeof(ki->p_comm), sizeof(p->p_comm))); 2272 strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename)); 2273 2274 ki->p_nlwps = p->p_nlwps; 2275 ki->p_realflag = ki->p_flag; 2276 2277 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2278 vm = p->p_vmspace; 2279 ki->p_vm_rssize = vm_resident_count(vm); 2280 ki->p_vm_tsize = vm->vm_tsize; 2281 ki->p_vm_dsize = vm->vm_dsize; 2282 ki->p_vm_ssize = vm->vm_ssize; 2283 ki->p_vm_vsize = atop(vm->vm_map.size); 2284 /* 2285 * Since the stack is initially mapped mostly with 2286 * PROT_NONE and grown as needed, adjust the "mapped size" 2287 * to skip the unused stack portion. 2288 */ 2289 ki->p_vm_msize = 2290 atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize; 2291 2292 /* Pick the primary (first) LWP */ 2293 l = proc_active_lwp(p); 2294 KASSERT(l != NULL); 2295 lwp_lock(l); 2296 ki->p_nrlwps = p->p_nrlwps; 2297 ki->p_forw = 0; 2298 ki->p_back = 0; 2299 ki->p_addr = PTRTOUINT64(l->l_addr); 2300 ki->p_stat = l->l_stat; 2301 ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag); 2302 ki->p_swtime = l->l_swtime; 2303 ki->p_slptime = l->l_slptime; 2304 if (l->l_stat == LSONPROC) 2305 ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags; 2306 else 2307 ki->p_schedflags = 0; 2308 ki->p_priority = lwp_eprio(l); 2309 ki->p_usrpri = l->l_priority; 2310 if (l->l_wchan) 2311 strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg)); 2312 ki->p_wchan = PTRTOUINT64(l->l_wchan); 2313 ki->p_cpuid = cpu_index(l->l_cpu); 2314 lwp_unlock(l); 2315 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 2316 /* This is hardly correct, but... */ 2317 sigplusset(&l->l_sigpend.sp_set, &ss1); 2318 sigplusset(&l->l_sigmask, &ss2); 2319 ki->p_cpticks += l->l_cpticks; 2320 ki->p_pctcpu += l->l_pctcpu; 2321 ki->p_estcpu += l->l_estcpu; 2322 } 2323 } 2324 sigplusset(&p->p_sigpend.sp_set, &ss2); 2325 memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t)); 2326 memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t)); 2327 2328 if (p->p_session != NULL) { 2329 ki->p_sid = p->p_session->s_sid; 2330 ki->p__pgid = p->p_pgrp->pg_id; 2331 if (p->p_session->s_ttyvp) 2332 ki->p_eflag |= EPROC_CTTY; 2333 if (SESS_LEADER(p)) 2334 ki->p_eflag |= EPROC_SLEADER; 2335 strncpy(ki->p_login, p->p_session->s_login, 2336 min(sizeof ki->p_login - 1, sizeof p->p_session->s_login)); 2337 ki->p_jobc = p->p_pgrp->pg_jobc; 2338 if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) { 2339 ki->p_tdev = tp->t_dev; 2340 ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2341 ki->p_tsess = PTRTOUINT64(tp->t_session); 2342 } else { 2343 ki->p_tdev = (int32_t)NODEV; 2344 } 2345 } 2346 2347 if (!P_ZOMBIE(p) && !zombie) { 2348 ki->p_uvalid = 1; 2349 ki->p_ustart_sec = p->p_stats->p_start.tv_sec; 2350 ki->p_ustart_usec = p->p_stats->p_start.tv_usec; 2351 2352 calcru(p, &ut, &st, NULL, &rt); 2353 ki->p_rtime_sec = rt.tv_sec; 2354 ki->p_rtime_usec = rt.tv_usec; 2355 ki->p_uutime_sec = ut.tv_sec; 2356 ki->p_uutime_usec = ut.tv_usec; 2357 ki->p_ustime_sec = st.tv_sec; 2358 ki->p_ustime_usec = st.tv_usec; 2359 2360 memcpy(&ru, &p->p_stats->p_ru, sizeof(ru)); 2361 ki->p_uru_nvcsw = 0; 2362 ki->p_uru_nivcsw = 0; 2363 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 2364 ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw); 2365 ki->p_uru_nivcsw += l2->l_nivcsw; 2366 ruadd(&ru, &l2->l_ru); 2367 } 2368 ki->p_uru_maxrss = ru.ru_maxrss; 2369 ki->p_uru_ixrss = ru.ru_ixrss; 2370 ki->p_uru_idrss = ru.ru_idrss; 2371 ki->p_uru_isrss = ru.ru_isrss; 2372 ki->p_uru_minflt = ru.ru_minflt; 2373 ki->p_uru_majflt = ru.ru_majflt; 2374 ki->p_uru_nswap = ru.ru_nswap; 2375 ki->p_uru_inblock = ru.ru_inblock; 2376 ki->p_uru_oublock = ru.ru_oublock; 2377 ki->p_uru_msgsnd = ru.ru_msgsnd; 2378 ki->p_uru_msgrcv = ru.ru_msgrcv; 2379 ki->p_uru_nsignals = ru.ru_nsignals; 2380 2381 timeradd(&p->p_stats->p_cru.ru_utime, 2382 &p->p_stats->p_cru.ru_stime, &ut); 2383 ki->p_uctime_sec = ut.tv_sec; 2384 ki->p_uctime_usec = ut.tv_usec; 2385 } 2386 } 2387
Indexes created Sun Sep 21 20:09:37 GMT 2025